Income distribution, financial liberalisations and banking stability: Theory and international evidence

Motivated by Rajan's study and existing economic reform theories (especially the lobby theory), we highlight income inequality as a driver of financial liberalisations. However, no evidence yet supports this idea. To verify this liberalisation‐induced‐by‐inequality hypothesis, the nexus of income inequality and financial liberalisations is formalised and empirically investigated in this paper. We utilise an ordered logit model with two‐way fixed effects and panel data from 91 countries to conduct the analyses. The baseline empirical results indicate that income inequality is a positive and significant determinant of financial liberalisations in aggregate and three component indicators. The results for three component indicators remain stable after a series of robustness checks, including mitigating the potential endogeneity via IV‐2SLS, Lewbel instrumental variable technique and difference within group, assessing the bias from the unobservable variables, considering the persistent and learning effects of liberalisations and the interaction effects, alternating the measures of income inequality, reducing business cycle, and using financial reforms rather financial liberalisation as independent variable. The results for aggregate liberalisations are slightly unstable in models with persistent and learning effects. Thus, liberalisation‐induced‐by‐inequality hypothesis is generally confirmed. In addition, the impacts of income inequality on financial liberalisations are widely heterogeneous across countries, the rate of change in income inequality and wealth inequality could also predict the occurrence of financial liberalisations, and the evidence that income inequality makes the poor easier access credits (i.e., Rajan's hypothesis) is found. Finally, the role is investigated of financial liberalisations in shaping banking stability, and we find that financial liberalisations can enhance growth and lower financial risks.